This paper proposes a method for planning a base position of mobile manipulators to perform picking tasks. By using our proposed method, we can find a base position of mobile manipulators where the manipulator can pick up an object placed at any position in a pallet. We first explain a method for solving the inverse kinematics taking the collision between a robot and a pallet into consideration. Then, we show a method for determining a base position. The effectiveness of the proposed method is confirmed by numerical examples.
This paper deals with a precise insertion processing (Peg-in-hole task) for speeding up and increasing efficiency in factory automation. When the clearance between the peg and the hole is narrower than the positioning accuracy of the robotic manipulator; for example, the former is 10[μm] and the latter is ±20[μm], it is near-impossible to insert the peg into the hole. Moreover in the case a robotic hand holds a plastic ring and inserts it into a shaft, the deformation of the ring makes the process more difficult. In order to solve these accuracy and deformation issues, we propose a novel insertion algorithm based on Passive Alignment Principle (PAP). When the ring is pressed against the shaft, the centers of them get together by the PAP effect so that the positioning error is corrected. Additionally, a robotic hand can relax the grip force without dropping the ring because of the friction of the contact points. So the ring deformation can be reduced. We conducted insertion experiments with the proposed algorithm. The experimental results validate our proposed method.
In recent years, in order to correspond to change of the industrial structure in the world, the industrial robot should be improved the flexibility and integration ability. RT-Middleware is suitable for the use in the coming industrial application that requires high-mix low-volume production because it has capability of rapid prototyping for robot system. However, robot system using RT-Middleware that have been published up to now be required the sophisticated or high cost technology for construction. In this study, the engineering sample by RT-Middleware to perform a pick and place task that is the basic behavior in the application of industrial robot are developed. To encourage the dissemination to industry of RT-Middleware, we have published the engineering sample with the manual for the introduction and operation.
One of the methods to design high speed robots is to adopt parallel mechanisms with driving force redundancy. However, such parallel mechanisms can have a problem in controlling them precisely because large feedback gains cannot be used if the large internal forces occur that might make the mechanisms broken. In this paper, we propose to introduce kinematical redundant degrees of freedom into the parallel mechanisms with driving force redundancy in order to solve this problem. The condition of the judgment is given in order to design such parallel mechanisms.
In automated assembly systems, it has been desired that to enable feeding incoming parts which is randomly piled in bins to those systems in terms of efficiency of parts storage and transportation. Previously, the number of parts type and equipments to align parts must be same. A dedicated supply device is called parts feeder. And for complex shaped parts, It had to rely to the manual labor. For robots, this is the Random-Bin-Picking problem itself, which is known as one of the classic challenges of robotics. The authors have developed a technology for bulk parts feeding with the flow work of generic robots which take out complex shaped parts used in the real assembly of electrical and electronic products from bulk parts boxes and then align them to parts pallets. The authors has built up a demonstration system using the developped technology, in which the first unit of the robot picks up real parts then the remaining three robots regrasp and align them. It was confirmed that parts of 11 kinds and more can be aligned to any position in the three-dimensional space in 6.45 seconds cycle in average, under 3 seconds interval in the best case.